eda-profiler 0.1.0

A lightweight package for detailed Exploratory Data Analysis on pandas DataFrames.

EDA Profiler

A lightweight yet comprehensive Python package for performing Exploratory Data Analysis (EDA) on a pandas DataFrame.

eda_profiler quickly generates a detailed profile report for each column in your dataset, providing essential statistics for data cleaning, feature engineering, and initial data understanding.

Features

• Calculates counts and percentages of missing values.
• Differentiates between numerical and categorical columns to provide relevant stats.
• For numerical columns, it computes:
  • Standard descriptive stats (mean, std, min, max).
  • A full range of percentiles (1%, 5%, 10%, 25%, 50%, 75%, 90%, 95%, 99%).
  • Distribution shape metrics: Skewness and Kurtosis.
  • Dispersion metrics: IQR and Coefficient of Variation.
  • Count of zero values.
• For categorical columns, it computes:
  • Cardinality (unique value count).
  • Mode (most frequent value), its frequency, and percentage.

Installation

You will be able to install this package via pip once it is published to PyPI. For now, you can install it locally from the project directory:

pip install .

How to Use

The package provides a single, easy-to-use function: profile_df.

import pandas as pd
import numpy as np
from eda_profiler import profile_df

# 1. Create a sample DataFrame
data = {
    'numeric_col': np.random.randn(100) * 100,
    'categorical_col': np.random.choice(['A', 'B', 'C'], 100, p=[0.6, 0.3, 0.1]),
    'mixed_col_with_nan': [1, 2, np.nan, 4, 5, 1, 2, np.nan] * 12 + [1,2,np.nan, 4]
}
df = pd.DataFrame(data)

# 2. Generate the EDA profile
eda_summary = profile_df(df)

# 3. Print the summary
# Transposing (.T) is often useful for readability
print(eda_summary.T)

Contributing

Contributions are welcome! Please feel free to submit a pull request or open an issue.

License

This project is licensed under the MIT License. See the LICENSE file for details.